I’ve been writing a fair bit about impact, and pool impact, values recently. Last week, on my blog post about creating power ratings, Michael asked me how to go about finding the best three horses using PIV’s (pool impact values), otherwise known as a/e ratios.
If you’d like to learn how to make PIV’s then you should check out this guide.
So we need to work out how to determine the best horses in a race based on PIV values.
Although we primarily use PIV’s to allow us to determine the importance of different factors for our handicapping, whether it’s a system, odds line or manual, we can also use it to help us find the strongest runners in the race.
Here is a screenshot showing the PIV’s over every type of flat racing:
I wouldn’t recommend you get stats on every single set of race conditions within flat racing, it’s far too broad a range of race conditions. But to show how to find the strongest horses in a race using PIV’s it will work fine 🙂
The first thing you notice is that the horses ranked 8 to 10 in the field for this factor have the strongest PIV’s.
The rating that I’m looking at is a consistency factor over the horses last ten starts so this actually makes sense.
What the PIV’s are telling us is that based on the odds a horse has, horses ranked 8 to 10 for this rating are proven to win more often than expected.
As you can imagine horses that are less consistent are going to be less bet on by the public, in this situation we’re seeing that the public are assigning to much importance to horses that don’t have a good consistency level in the last ten starts.
This follows common thinking that you should look at the last six form figures and look for runners that have performed well. Without looking at them in more detail punters are actually under-valuing the runners who may be less consistent but have the ability to contend.
But I’m getting off track!
The focus of this post is how do we find the strongest runners based on this information. And there are a couple of ways I’m going to recommend.
What you need to do is make a list of horses for the race you’re interested in.
As you can see, Charming Lad is ranked number 1 and we look at what PIV these runners have received historically, and it’s 0.97. So Charming Lad would be given 0.97.
Brunette’Sonly is ranked number two and these horses have historically received a PIV for this factor of 1.01 so this is what this horse gets. We do this for each of the horses.
Now if you were using a power rating then you could just consider taking the strongest runners, but this is a factor that only measures a horses consistency. That’s not exactly an good overall assessment of the horses ability in the race.
We may have another two or more factors that we want to add PIV’s for. For example…
I’ve just put some random numbers into the second column to demonstrate what we need to do combine the figures when we have more than one per horse. If you ask a statistician then they’ll tell you that neither are perfect. There issues with factors being correlated and other things. And the truth is…
…they’re right!
But all we want is something that can be done quickly and as effectively as possible because the final decision is going to be made by us not an automated machine process. And that makes all the difference.
There are two ways I recommend you combine the PIV’s.
Approach 1
You can multiply the PIV’s together to give you….
The Multiplied column can then be used to determine which horses are most likely to contend in the race.
Approach 2
The alternative approach is to use the Reynolds Ranking method which I wrote about here. This involves changing the PIV’s into rankings, i.e. the horse with the highest PIV is ranked 1, the second highest is ranked 2 etc…
Once you’ve done that you multiple the two rank columns together…
The lower the figure in the Multiplied column then the stronger the horse.
In this example Rivermouth and Dr Dreamy would be the strongest with Charming Lad looking to be the next strongest.
However remember that this includes very generic data from just one factor combined with some random numbers to demonstrate how the process works.
Before we finish up I would like to give you the heads-up on something that will not only make a difference to how you use this approach, but how you use all ratings:
Never just consider the horses raw ratings or the top rated horses. Always consider how big the gaps are between the runners in your decision about which horses should be considered a threat.
